One way of describing the distribution of income is to compare the share of total income enjoyed by different groups of the distribution. The larger the income share of the richest, and the smaller the income share of the poorest, so the greater the inequality in income distribution. In the EU in 2013, the income share of the richest 10% of the population was largest in Cyprus (where the top 10% had 29% of total income), while in Bulgaria, Lithuania, Latvia and Portugal, the top 10% had 26-27% of total income. The share was smallest in Sweden (around 20%) and the figure was only marginally higher in Slovenia and Belgium (Figure 1).
Figure 1: The income share of the top 1%, 5% and 10% of the population in EU Member States, 2013 income year
There is more variability among EU Member States in the income share of the top 5% and the top 1%. However, given the relatively small sample size surveyed, this variability may be a result of data problems. The country ranking according to the share of the top 5% is almost identical to the ranking based on the share of the top 10%. The country rankings according to the income share of the richest 10% and of the richest 1% are similar but not identical (the correlation coefficient is 0.7) (Table 1 ).
The bottom 10% of the population had between 2% and 4% of total income in EU Member States in 2013 (Figure 2). The smallest share was found in Romania, Bulgaria, Spain and Greece (between 2.0 and 2.3%), while the largest was in the Czech Republic, Finland and the Netherlands, where the income share of the poorest decile exceeded 4%. The rank order of countries according to the bottom 10% and the bottom 5% is again similar, while the correlation of the rankings according to the bottom 10% and bottom 1% is somewhat higher than in case of the top of the income distribution (Table 2 ).
Figure 2: The income share of the bottom 1%, 5% and 10% of the population in EU Member States, 2013 income year
The most widely used measure for income inequality is the Gini coefficient . The maximum possible value of the Gini coefficient is 1 (when one individual has all the income in a country), while the lowest value is 0 (when everyone has the same income). In Figure 3, the Gini coefficient is shown together with 95% confidence intervals. In the EU, the value of the Gini coefficient in 2013 ranged from 0.25 (in Slovenia, Sweden and the Czech Republic) to 0.35 (in Bulgaria, Cyprus and the Baltic states). Other countries at the top of the ranking were Portugal, Romania, Greece and Spain with Gini indices around 0.34. At the bottom of the country ranking, Finland, the Netherlands, Belgium and Slovakia have Ginis that are only slightly higher than Slovenia's (between 0.25 and 0.26). Other countries can be broadly divided into two groups, with Croatia, Ireland, Poland, the UK and Italy having Ginis of between 0.30 and 0.33, and other EU15 countries together with Malta and Hungary having values of between 0.25 and 0.30.
Figure 3: Gini coefficients and 95% confidence intervals for disposable household income in EU Member States, 2013 income year
The level of GDP per head and income equality are positively correlated among both the EU15 and EU13 Member States (Figure 4). Among the EU13 Member States, Croatia, Hungary, Slovakia, the Czech Republic and Slovenia have lower levels of inequality than would be expected given their level of GDP per head (i.e. they fall below the line indicating the average relationship in the Figure). Estonia, Latvia and Lithuania on the other hand have higher levels of inequality. Among the EU15 countries Finland, Belgium and Sweden have a lower degree of inequality than their GDP per head would suggest, while Ireland and Germany, Spain and Cyprus have somewhat higher degrees of inequality.
Figure 4: The relationship between income equality and GDP per head, 2013 income year
In addition to the Gini, several other inequality measures have been developed by researchers. Given the possibility of data problems for very low and very high incomes and also the sensitivity of some of the measures to extreme values, a bottom and top coding procedure was applied in the calculations reported here (see the section on Measurement and Methodology).
The other measure of income inequality commonly used in the EU as an indicator is the ratio of the income share of the richest 20% of the population to the share of the poorest 20% (the S80/S20 ratio). The higher the value of this, the more unequal is the distribution. In the Czech Republic, Finland and the Netherlands the top 20% has approximately 3.5 times the income of the bottom 20%; in Romania and Bulgaria they have incomes almost 7 times higher than the poorest fifth of the population (Figure 5).
Other measures include the squared coefficient of variation (SCV), the mean logarithmic deviation (MLD), the Theil-index and the Atkinson-indices. Most of these show similar rankings (Table 3 ) compared to the Gini ranking: the correlation coefficient between the ranking and the ranking according to the Gini index is close to one. Exceptions are inequality measures that are more sensitive at the tails (or the two extreme ends) of the income distribution than is the Gini coefficient. E.g. the SCV tends to be sensitive to high incomes. It is noteworthy that Croatia, Estonia, Spain, Poland and Latvia have a lower inequality ranking (meaning lower inequality compared to other countries) according to this measure than according to the Gini, while the ranking of Denmark, France, Germany, Romania, Greece and Cyprus increases by three or more positions. The Atkinson (ε=2) index tends to be relatively sensitive to lower incomes. According to this measure, Sweden, Austria, Slovakia, Spain, Romania and Italy have a higher degree of inequality relative to other countries than is indicated by the Gini, while six countries, most importantly Cyprus and Lithuania have a lower degree of inequality.
 The Gini coefficient measures the area between the Lorenz curve (which plots the cumulative distribution of income - the percentage of income going to a given percentage of the population, when the latter is ranked according to income levels) and the line of complete equality (the 45-degree line, where a given percentage of income goes to the same percentage of population).